Virtual reality (VR) is an often-used instrument in sports science research and practical training. However, VR studies with experienced athletes and sports specific tasks are rare. Furthermore, the transfer from interventions in VR into reality is even less investigated. It is possible to analyze benefits of VR using in-situ studies comparing human behavior in VR with reality. If no differences occur in the human behavior, then VR would be appropriate for interventions to improve athletes’ performance. Therefore, we let seven karate athletes respond each to ten attacks of a real attacker (reality) and a virtual attacker (VR using a Head Mounted Display) and compared the parameters “response quality” and “attack recognition” under both conditions. As attacks we chose Gyaku-Zuki (reverse punch, GZ) and Kizami-Zuki (attack with the front arm, KZ). ANOVAs and sign tests showed isolated cases of significant differences between both conditions: response quality in KZ, and attack recognition for 150ms in GZ, all p0.05). The remaining comparisons showed no significant differences (p>0.05). We conclude that further research is needed but the results of the present pilot study are promising to assume that VR is suitable for applications because similar performance outcome in reality and VR were obtained.